arX

iv:0

904.

2379

v1 [

astr

o-ph

.GA

] 1

5 A

pr 2

009

Mon. Not. R. Astron. Soc. 000, 1–9 (2009) Printed 1 November 2018 (MN LATEX style file v2.2)

Major Dry-Mergers In Early-Type Brightest Cluster

Galaxies

F. S. Liu1,2⋆, Shude Mao3, Z. G. Deng4, X. Y. Xia5, and Z. L. Wen21College of Physics Science and Technology, Shenyang Normal University, Shenyang, 110034, P.R.China2National Astronomical Observatories, Chinese Academy of Sciences, A20 Datun Road, Beijing, 100012, P.R.China3Jodrell Bank Centre for Astrophysics, Alan Turing Building, University of Manchester, Manchester, M13 9PL, UK4Graduate University of Chinese Academy of Sciences, Beijing, 100049, P.R.China5Tianjin Astrophysics Center, Tianjin Normal University, Tianjin, 300074, P.R.China

Accepted 2009 April 14, Received 2009 April 5; in original form 2009 January 27

ABSTRACT

We search for ongoing major dry-mergers in a well selected sample of local Bright-est Cluster Galaxies (BCGs) from the C4 cluster catalogue. 18 out of 515 early-typeBCGs with redshift between 0.03 and 0.12 are found to be in major dry-mergers,which are selected as pairs (or triples) with r-band magnitude difference δmr < 1.5and projected separation rp < 30 kpc, and showing signatures of interaction inthe form of significant asymmetry in residual images. We find that the fraction ofBCGs in major dry-mergers increases with the richness of the clusters, consistentwith the fact that richer clusters usually have more massive (or luminous) BCGs.We estimate that present-day early-type BCGs may have experienced on average∼ 0.6 (tmerge/0.3Gyr)−1 major dry-mergers and through this process increases theirluminosity (mass) by 15% (tmerge/0.3Gyr)−1 (fmass/0.5) on average since z = 0.7,where tmerge is the merging timescale and fmass is the mean mass fraction of compan-ion galaxies added to the central ones. We also find that major dry-mergers do notseem to elevate radio activities in BCGs. Our study shows that major dry-mergersinvolving BCGs in clusters of galaxies are not rare in the local Universe, and they arean important channel for the formation and evolution of BCGs.

Key words: galaxies: elliptical and lenticular, cD - galaxies: clusters: general - galax-ies: photometry

1 INTRODUCTION

The Brightest Cluster Galaxies (BCGs) are at the most lu-minous and massive end of galaxy population. They are usu-ally located at or close to the centres of dense clusters ofgalaxies based on the X-ray observations or gravitationallensing observations (e.g., Jones & Forman 1984; Smith etal. 2005). Most of them are dominated by old stars with-out prominent ongoing star formation. They are much moreluminous than other galaxies in the clusters, and have dis-tinct surface brightness profiles from other galaxies (see Vale& Ostriker 2008 and references therein). And yet the lumi-nosities of BCGs appear to have rather small scatters. As aresult, they have been proposed as a distance indicator (e.g.Postman & Lauer 1995). Because of the unusual properties

⋆ E-mail: lfs@bao.ac.cn

of BCGs, their formation and evolution has long attractedparticular attention.

Galactic cannibalism as a way to build up the BCGswas proposed more than three decades ago (e.g., Ostriker &Tremaine 1975; White 1976; Vale & Ostriker 2008). How-ever, as pointed out by several authors (e.g., Merritt 1985;Lauer 1988; Tremaine 1990), cannibalism of satellite galaxiesmay not be enough to account for the luminosity growth ofmassive BCGs due to the small luminosity ingestion rate. Ifthis is correct, then other mechanisms, such as major merg-ers may play an important role (Lin & Mohr 2004; Vale &Ostriker 2008). Since most central galaxies in clusters havelittle cold gas, the mergers are not expected to trigger ma-jor episodes of star formation and thus they are likely drymergers.

Recent studies from both numerical simulations and ob-servations indicate that a large part of stellar mass in lu-minous early-type galaxies form before redshift z ∼ 1 − 2,

c© 2009 RAS

2 F. S. Liu et al.

and later dry mergers play an important role in their stel-lar mass assembly (Gao et al. 2004; De Lucia & Blaizot2007; van Dokkum et al. 1999; Tran et al. 2005). Ruszkowski& Springel (2009) investigated the effect of dry-mergers onthe scaling relations of BCGs in simulations. Observationalstudies on the properties of luminous early-type galaxiesprovide evidence that they have experienced dry-mergers(e.g., Lauer et al. 2007; Bernardi et al. 2007; von der Lindenet al. 2007; Liu et al. 2008, hereafter Paper I). Theoreticalinvestigations suggest that half the mass of a typical BCGmay be assembled via accretion of smaller galaxies, i.e., byminor mergers at z < 0.5 (e.g., Gao et al. 2004; De Lu-cia & Blaizot 2007). Recently, Bernardi (2009) studied theirevolutions in the size- and velocity dispersion-stellar masscorrelations since z < 0.3, and concluded that early-typeBCGs grow from many dry minor mergers (see Discussion).However, Whiley et al. (2008) using the ESO Distant Clus-ter Survey, found little evolution of BCGs since z ∼ 1, atleast within a metric circular aperture of 37 kpc. A consensusabout the evolution of BCGs is yet to emerge.

Recent modelling suggests that major dry-mergers playa significant role in forming luminous, intermediate-mass,early-type galaxies (e.g. Naab et al. 2006) at redshift z 6 0.5.Whether the same scenario applies to the much more lumi-nous BCGs, and more generally how the occurrence of ma-jor dry-mergers depends on the environments, is still some-what controversial. Merritt (1985) argued that massive ma-jor mergers are more likely to occur in large groups than inmassive clusters, since the merger rate is a steeply decliningfunction of the velocity dispersions of a virilized system.

However, recent observational studies indicate majordry-mergers do occur in groups and clusters of galaxies. Forexample, Mulchaey et al. (2006) and Jeltema et al. (2007)reported some examples of dry-mergers involving centralgalaxies in intermediate-redshift groups. Tran et al. (2008)reported an observational analysis of supergroup SG 1120-1202 at z ∼ 0.37, which is expected to merge and form acluster with mass comparable to Coma. They argued thatthe group environment is critical for the process of majordry-mergers. Rines et al. (2007) reported a very massivecluster (CL0958+4702) at moderate redshift (z = 0.39), inwhich a major dry-merger is ongoing to build up a BCG.Using the Sloan Digital Sky Survey (SDSS) data, McIntoshet al. (2008) identified 38 major mergers of red galaxies from845 groups of galaxies at z 6 0.12 identified using the halo-based group finder of Yang et al. (2005), and showed thatcentres of massive groups are the preferred environment formajor dry-mergers and mass assembly of massive early-typegalaxies.

All these studies show that major dry-mergers are im-portant in the formation of BCGs at moderate redshifts. Ourown visual inspections of 249 merging pairs identified fromluminous early-type galaxies (Wen, Liu & Han 2009) alsohint that some of these mergers involve BCGs. The purposeof this work is to perform a search for major dry-mergersin a well-selected local early-type BCG sample with a quan-titative method and then estimate their major dry-mergerand luminosity (mass) increase rates. We also investigatethe dependence of the merger fraction as a function of therichness of the clusters.

The structure of the paper is as follows. In §2 and §3we describe our sample and identifications of major dry-

mergers involving BCGs. We present our main results in §4and finish with a summary and discussion in §5. Through-out this paper we adopt a cosmology with a matter densityparameter Ωm = 0.3, a cosmological constant ΩΛ = 0.7,and a Hubble constant of H0 = 72 kms−1Mpc−1, i.e.,h = H0/(100 kms−1Mpc−1) = 0.72.

2 THE SAMPLE

The C4 cluster catalogue (Miller et al. 2005) was constructedfrom the SDSS spectroscopic data in the parameter space ofposition, redshift, and color. The C4 algorithm identifies theBCG from the spectroscopic catalogue within 0.5h−1 Mpcof the central galaxy at the peak of the C4 density field(“mean” galaxy). The early (DR2) version of the C4 cat-alogue incorporates the SDSS photometric catalogue to se-lect the brightest cluster galaxy within 1h−1 Mpc in order toavoid missing BCGs in spectroscopic data due to fiber col-lisions. The catalogue is relatively complete in the redshiftrange 0.03 6 z 6 0.12, including 98% of X-ray-identifiedclusters and 90% of Abell clusters in the same region (Milleret al. 2005). We thus limit our data within this range of red-shift, obtaining 643 clusters after we reject 16 duplicates(Bernardi et al. 2007).

However, as pointed out by several previous works (e.g.,Bernardi et al. 2007; von der Linden et al. 2007), some brightstars have been mis-classified as BCGs, and about 1/4 ofBCGs show late-type features (e.g., distinct spiral arms,dust features). We thus re-checked each cluster and iden-tified the brightest galaxy within 1h−1 Mpc of the clustercenter as the BCG. We then select a sample of early-typeBCGs from this catalogue by excluding artifact stars andlate-type galaxies. Furthermore we require the BCGs shouldbe in clusters with richness 1 larger than 10, and have colourtypical of early-type galaxies, g − r > 0.7 (Strateva et al.2001) in the SDSS model magnitudes. In the end we obtaina sample of 515 early-type BCGs in the local Universe. Fig-ure 1 shows the distributions of redshifts of the clusters andg − r colours of BCGs, which has taken into account theextinction by the SDSS, and k-correction to z = 0.1 (Blan-ton & Roweis 2007). The cluster redshifts range from 0.03to 0.12, with a maximum around 0.08; the g − r colours ofBCGs peak around 0.9 with a small RMS scatter of around0.06, consistent with their stellar populations being quiteold and uniform.

3 IDENTIFICATIONS OF MAJOR

DRY-MERGERS INVOLVING BCGS

Dry-mergers between gas-poor early-type galaxies are gen-erally more difficult to identify than those mergers involv-ing late-type galaxies (Naab et al. 2006). The latter oftendevelop prominent tidal tails dotted with star-forming re-gions. However, dry-mergers have more subtle features, suchas broad “fans” due to ejected stars, asymmetries in in-ner isophotes and/or sometimes diffuse tails (Rix & White1989; Combes et al. 1995). While major dry-mergers usually

1 The richness is defined as the number of member galaxies within1h−1 Mpc of the central galaxy.

c© 2009 RAS, MNRAS 000, 1–9

Major Dry-Mergers In BCGs 3

Figure 1. The top panel shows the redshift distribution of red-shift zcl for the 515 C4 clusters while the bottom panel shows thecolour g − r distribution of their early-type BCGs.

have more prominent morphological signatures than minordry-mergers (Le Fevre et al. 2000), they still exhibit muchweaker merging features than wet-mergers.

We adopt the following criteria to select candidatesof ongoing major dry-mergers involving the BCG in oursample. First, galaxies with projected distance from centralBCG must satisfy rp < 30 kpc. We search for close pairs bothautomatically and visually. In the automated search, a lowerlimit of the projected distance, 7 kpc, is imposed, which cor-responds to a separation of 3′′ at z = 0.12. This restrictionis necessary because the photometry becomes unreliable forgalaxies with smaller angular separation (McIntosh et al.2008; Wen, Liu & Han 2009). Pairs or triples with separa-tions smaller than 7 kpc are searched visually. Second, theyshould have magnitude difference with their correspondingcentral BCGs < 2 magnitudes in the SDSS r−band modelmagnitude. The magnitude difference applied here is slightlylarger than the value (1.5) that will be eventually used toselect major-mergers. The reason is that the SDSS magni-tudes for BCGs are somewhat inaccurate due to the problemof sky background subtraction (Paper I). Third, both galax-ies in a pair must have g − r > 0.7 to ensure the merginggalaxies are early-type (Strateva et al. 2001). In total, thereare 49 apparent close pair candidates at 7 < rp < 30 kpcsatisfying these three conditions (pairs with rp < 7 kpc arediscussed below).

Most previous studies impose an additional requirementof a small difference in the line-of-sight velocities δvr (e.g.,

Figure 2. The r−band residual magnitudes mres versus theasymmetry factor A. The black points are the pairs in Wen etal. (2009, Fig. 5 in their paper). The 36 merger candidates with7 < rp < 30 kpc are shown with red circles. Four additional merg-ers with rp < 7 kpc are shown with red triangles. The horizontaland vertical dotted lines illustrate the criteria of Wen et al. e.g.,mres < 20mag and A > 0.45 used to select major mergers. Thehistograms for A and mres for the Wen et al. (2009) sample areshown at the right and bottom for major-mergers with stronginteractions (solid histogram) and those classified as weak inter-actions or due to chance alignment (dashed histogram).

δvr < 500 kms−1) besides a small projected distance rp toavoid chance alignment. However, SDSS spectroscopic sur-vey is severely incomplete on very small angular scales due tofiber collision (McIntosh et al. 2008; Wen, Liu & Han 2009).Hence it is not possible to apply this criterion to the SDSSdata. An alternative method is needed to identify mergingsystems from the projected pairs and triples.

Although, as mentioned above, the interaction features(e.g., broad plumes at the outskirts, short tidal tails, bridgesor asymmetries in inner isophotes) are weak in mergersof early-type galaxies, they can nevertheless be seen moreclearly from the residual image after we subtract a smoothand symmetric model for each galaxy in an image (Bell etal. 2006; McIntosh et al. 2008; Wen, Liu & Han 2009). Sucha procedure has been described in detail by Wen, Liu &Han (2009). For completeness, we repeat the essentials ofthe method below.

First, we extract the SDSS r−band images of all 49close pairs and triples. The sky backgrounds are subtractedprecisely from the corrected frames with the method of Pa-per I. Briefly, the background is subtracted as follows: allthe objects including target sources in the corrected framewere firstly masked to obtain a background-only image.

c© 2009 RAS, MNRAS 000, 1–9

4 F. S. Liu et al.

Second-order Legendre polynomials are then used to fit thesmoothed background-only image to obtain an accurate skybackground model. After background subtraction, the GAL-FIT package (Peng et al. 2002) is then used to construct asmooth symmetric Sersic (Sersic 1968) model for every early-type galaxy in the projected pair or triple. Stars and faintergalaxies with distances closer than 2R90 from the centre ofBCGs (Here R90 is the radius containing 90% of Petrosianflux from the SDSS catalogue) are modelled with a double-Gaussian point-spread-function (Stoughton et al. 2002) andSersic function respectively. Objects outside 2R90 in the ex-tracted images are masked. All models are convolved withthe point-spread-function; we then obtain the best modelby minimizing the χ2 with the sky-subtracted image of un-masked pixels.

A fitted magnitude for each target galaxy and a resid-ual image for each pair are obtained afterwards. Notice thatthe fitted magnitude is superior because it can separate theflux in the overlapping region of pairs and correct the over-estimate in the sky background in the SDSS pipeline (PaperI). Wen, Liu & Han (2009) improved the method of Conseliceet al. (2000) to calculate the asymmetry factor for residualimages of galaxy pairs. Essentially it measures the differencebetween pixels and those symmetric pixels with respect tothe centre and major axis of each galaxy (see eqs. 8-10 intheir paper). The pixels in overlapping regions are treatedseparately (see their Figure 4). The asymmetry factor A ∼ 0is for a galaxy pair without any interaction feature ideally. Alarge A means a stronger asymmetric interaction. In orderto find an efficient criterion, they compared the asymme-try factors of two test samples of residual images selectedthrough careful visual inspections: a sample with distinctinteraction features and another almost without. Both arenot contaminated by other objects within 2R90 of targetgalaxies. The test showed that a close pair can be classi-fied efficiently as a merging system with distinct interactionfeatures when the residual images have an r−band magni-tude rres < 20 mag and an asymmetry factor A > 0.45.In this work, we performed further extensive tests and findthat both the asymmetry parameter A > 0.5 and residualmagnitude rres < 19.5 can select strongly interacting merg-ers efficiently (see the two histograms in Figure 2). Thesecriteria are only slightly different from those used by Wen,Liu & Han (2009). The criteria of rres < 19.5 mag is almostthe same as requiring the flux ratio of the residual image tothe initial image to be larger than 2% (fres/fini > 2%, seethe last column in Table 1).

If we limit major mergers as having a fitted magnitudedifference δmr < 1.5, we find that 36 out of 49 close pairsor triples satisfy 7 < rp < 30 kpc. Furthermore, 14 out of36 have residuals rres < 19.5 mag and asymmetry factorsA > 0.5 (see Figure 2). Others may be weak interacting sys-tems or pairs in chance alignment; two examples with smallresiduals are shown in Figure 4. We added 4 mergers (C1035,C1176, C3150, C3157) with rp < 7 kpc (e.g., mergers withclose double nuclei) identified visually – these systems havefiber magnitude difference δmr < 1.0, and can be classifiedas merging systems if we apply the same criteria as merg-ers at 7 < rp < 30 kpc. The r−band observational imagesand corresponding residual images for these 18 mergers in-volving BCGs are shown in Figure 3. Their correspondingparameters are listed in Table 1.

Figure 4. Two typical example galaxies which after subtractionshow small residuals and small asymmetry factors. They are likelypairs produced by either weakly interacting with each other orchance alignment. Each image is 80′′ × 80′′.

Bernardi et al. (2003) derived the luminosity functionof early-type galaxies. In the r-band, the magnitude for anL∗ galaxy (derived from a Gaussian fit) is M∗ = −21.09(taking h = 0.72 instead of their h = 0.70) . The BCGsin our sample have luminosities ranging from L∗ to roughly∼ 10L∗, with a mean luminosity of ∼ 5L∗. For C4 2089, themerging pair has luminosities 11.2L∗ and 5.6L∗. Similarly,for C4 1055, the triple merging galaxies have luminosities7.9L∗, 6.6L∗ and 0.46L∗; the two most luminous componentshave identical radial velocities, and thus there is little doubtthat they are physically associated. Mergers appear to occureven among the most luminous galaxies in clusters (see §4.2).If all the stars coalesce to form a single giant galaxy, thenthe luminosities will be as high as 16.8L∗ and 15L∗; suchgalaxies are rarely seen. This suggests that a substantialamount of stars of the companion are tidally stripped toform an “envelope” around the central galaxy, a point wereturn to in §4.1.

4 RESULTS

4.1 Merger rate and stellar assembly of BCGs

We have identified 18 (∼3.5% of the total) major dry merg-ing pairs (or triples) involving the BCG in 515 C4 clustersat 0.03 6 z 6 0.12 . It has usually been assumed that physi-cally bound pairs will merge due to dynamical friction. Themerger timescale is, however, uncertain since it depends ona number of factors, including the separation, relative veloc-ity and mass ratios of galaxies. Kitzbichler & White (2008)noticed that previous studies on merger rates based on pairstatistics have yielded a wide variety of results (see theirpaper for a review). This diversity can be attributed to dif-ferences in the pair definition and in the timescales adopted.They thus used the Millennium Simulation to calibrate themerging timescale as a function of the stellar mass and pro-jected separation. They find an average merging timescale(appropriate for our stellar mass and the Hubble constant):

tmerge = 2.2Gyrrp

50 kpc

„

M∗

5.6 × 1010M⊙

«−0.3“

1 +z

8

”

, (1)

where M∗ is the total stellar mass, and z is the (cluster) red-shift. We take a mass-to-light ratio of M∗/Lr ≈ 5M⊙/L⊙

(Rines et al. 2007; Patton et al. 2000), yielding a mean totalstellar mass of M∗ ∼ 8.5 × 1011M⊙ for our massive merg-ers. We estimate the merger timescale for each merger bythis formula. We find the merger timescale of our mergersranges from 0.04 to 0.55Gyr with a mean value of 0.3Gyrand a standard deviation of 0.17Gyr (see Column 11 in Ta-ble 1). For comparison, if we apply the formula of Masjedi et

c© 2009 RAS, MNRAS 000, 1–9

Major Dry-Mergers In BCGs 5

Figure 3. The colour images and corresponding residual images in the r−band for the 18 identified major mergers. Each image is80′′ × 80′′. The SDSS-C4 name for the DR2 version is marked at the top left corner of each colour image.

al. (2006) to calculate the dynamical timescale of our merg-ers, we obtain a similar timescale of ∼ 0.3Gyr. However, asomewhat shorter dynamical friction timescale of ∼ 0.1Gyris obtained if we apply the formula of Patton et al. (2000),although the dynamical friction formula may be less appli-cable to major mergers (Binney & Tremaine 1987). We alsonote that Bell et al. (2006) obtained a shorter timescale of∼ 0.15Gyr from simulations for massive early-type merg-ers. It seems that the average merging timescale is likelyuncertain by a factor of ∼ 2.

To evaluate the number of mergers BCGs has experi-enced since (for example) z = 0.7, we need to understandhow the dry-merger fraction evolves as a function of redshift.Parameterizing the evolution of pair fraction in the form of(1+z)m, several previous studies (e.g., Le Fevre et al. 2000;Patton et al. 2002; Conselice et al. 2003; Lin et al. 2004;De Propris et al. 2005; Kartaltepe et al. 2007; Lotz et al.2008) derive a positive slope in the range of m ∼ 0 − 3 forall mergers (involving dry or wet mergers). Lin et al. (2008),for the first time, estimated the merger fraction of red galax-ies (the majority of them should be dry-mergers) decreaseswith redshift slightly with m = −0.92± 0.59. This suggeststhat dry-mergers may preferentially occur in the local Uni-verse. In contrast, a recent work by Khochfar & Silk (2008)suggests a nearly constant dry-merger rate at z 6 1.

If we adopt an average merger timescale, and assumea constant fraction (18/515 ∼ 3.5%) of BCGs are in ma-jor dry-mergers over the 6.3 Gyr interval since redshift

0.7, then a present-day early-type BCG will have experi-enced 6.3/tmerge × 18/515 ∼ 0.7 (tmerge/0.3Gyr)−1 majordry-mergers. On the other hand, if the fraction evolves as(1+z)−0.92±0.59 (Lin et al. 2008), then the number of merg-ers changes to ∼ 0.5(tmerge/0.3Gyr)−1. Below we adopt anaverage number of mergers of these two evolution scenar-ios, 0.6 (tmerge/0.3Gyr)−1. In any case, the uncertainty inthe merger timescale is much larger than that due to theevolution of the merger rate.

Our merger pairs/triples have an average magnitudedifference of ∼ 0.7, which corresponds to a ∼ 1 : 2 lumi-nosity ratio. In the merging process, a substantial fractionof the companion galaxy may be tidally stripped, and sonot all the mass of the companion galaxy will be added tothe main one (Yang et al. 2009). We assume a fraction offmass of the companion galaxy is accreted to the primarygalaxy. Thus in each merger, the central galaxy increasesits luminosity (or mass) by 25%(fmass/0.5). It follows thata present-day BCG has on average increased its luminos-ity by 15% (tmerge/0.3Gyr)−1(fmass/0.5) from z = 0.7 at arate of 2.5% (tmerge/0.3Gyr)−1(fmass/0.5) per Gyr. Thus anon-negligible fraction of the stellar mass of a present-dayearly-type BCG is assembled via major dry-mergers sincez = 0.7.

c© 2009 RAS, MNRAS 000, 1–9

6 F. S. Liu et al.

Table 1. Basic parameters for the 18 identified major dry-mergers involving BCGs

C4 ID zmg Mtot R.A.(J2000) Dec.(J2000) zsp rp δvr mr Mr tmerge A mres fres/fini(mag) ( kpc) ( km s−1) (mag) (mag) (Gyr) (mag) (%)

(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14)

1000 0.0864 -23.36 202.543025 -2.105012 0.0866 15.42 -22.62 0.32 0.75 17.98 5.2202.545566 -2.103784 0.0863 15.94 90 15.44 -22.60

1004 0.0814 -23.30 149.717402 1.059187 0.0814 14.90 -22.99 0.27 0.78 19.04 2.5149.719361 1.060678 – 13.18 – 16.09 -21.80

1011 0.0903 -23.91 227.107326 -0.266266 0.0903 14.87 -23.27 0.39 2.71 17.79 5.1227.104246 -0.268636 – 22.92 – 15.51 -22.63227.103726 -0.270214 – 31.50 – 16.38 -21.76

1026 0.0908 -24.16 191.926938 -0.137254 – 14.50 -23.65 0.28 1.15 18.36 2.3191.927005 -0.140256 0.0908 17.78 – 15.06 -23.09

1035∗ 0.0444 -22.95 175.872249 -1.667968 0.0446 16.11a – 0.04 0.69 18.46 2.1175.872512 -1.668436 0.0442 1.64 120 16.16a –

1055 0.0836 -24.03 202.795956 -1.727285 0.0836 14.61 -23.34 0.28 0.62 18.39 2.1202.795126 -1.730259 0.0836 16.99 0 14.81 -23.14202.796700 -1.725844 – 8.93 – 17.70 -20.25

1060 0.1170 -23.29 212.497855 -1.539655 – 15.87 -22.91 0.17 3.49 19.09 4.2212.497645 -1.538576 0.1170 8.13 – 16.81 -21.97

1176∗ 0.0737 -23.90 189.734808 6.158386 0.0738 17.06a – 0.11 1.53 17.34 3.3189.734648 6.157096 0.0736 6.37 60 17.62a –

1304 0.0986 -22.69 148.914494 1.596777 0.0987 16.12 -22.23 0.50 0.87 18.55 7.4148.917693 1.597327 0.0986 20.72 30 16.84 -21.51

1364 0.0708 -22.75 154.700071 0.385141 0.0710 15.06 -22.49 0.50 0.82 18.54 4.9154.701081 0.380757 0.0707 21.27 90 16.49 -21.06

2049 0.0751 -23.89 16.841087 14.273220 – 14.05 -23.64 0.13 0.77 18.04 2.216.840364 14.274536 0.0751 7.44 – 15.52 -22.17

2089 0.0652 -24.15 59.582654 -5.538873 0.0652 13.65 -23.71 0.27 0.86 17.92 5.159.583713 -5.535151 – 16.95 – 14.40 -22.96

2179 0.0938 -22.24 322.027369 11.405484 – 16.71 -21.53 0.47 0.57 18.43 10.6322.027823 11.408263 0.0938 17.17 – 16.80 -21.44

3059 0.0406 -22.47 122.148487 38.914504 – 14.26 -22.00 0.43 1.12 17.62 4.4122.140730 38.914841 0.0406 16.97 – 14.94 -21.32

3150∗ 0.0456 -22.05 134.299844 53.470904 0.0456 16.43a – 0.06 0.71 17.72 4.8134.301020 53.470989 – 2.21 – 17.09a –

3157∗ 0.0635 -24.10 176.404931 64.511485 0.0632 16.37a – 0.08 1.23 18.03 2.3176.406011 64.512554 0.0638 4.97 180 16.89a –

3166 0.1112 -24.04 192.035328 64.036916 0.1112 14.93 -23.72 0.49 0.53 19.29 3.3192.044829 64.037322 – 29.63 – 16.11 -22.54

3167 0.0915 -23.36 256.010966 33.868847 0.0915 15.88 -22.29 0.55 1.92 19.01 3.5256.014668 33.872302 0.0919 27.66 120 15.94 -22.23256.008867 33.870576 0.0911 14.55 90 16.25 -21.92

Note: Col:(1). The C4 cluster ID (DR2) for the Clusters. The four visually-selected ones (indicated by ∗) have rp < 7 kpc (see §3).Col:(2). Redshifts of identified merging pairs (or triples). Col:(3). Total absolute magnitude of mergers in the SDSS r−band. Col:(4).R.A. (J2000) of the component galaxy in a merger. Col:(5). Dec. (J2000) of the component galaxy in a merger. Col:(6). Spectroscopicredshift of the component galaxy in a merger. Col:(7). The projected distance, rp. Col:(8). The line-of-sight velocity difference, δvr, ifavailable. Col:(9). Extinction-corrected r−band fitted apparent magnitude of component galaxy in a merger. For four mergers withrp < 7 kpc (marked with ‘a’), we compare their fiber magnitudes. Col:(10). r−band absolute magnitude of component galaxy in amerger, corrected for extinction (by SDSS) and the k-correction (using the KCORRECT algorithm of Blanton & Roweis 2007). Col:(11).The merger timescale estimated by the formula of Kitzbichler & White (2008). Col:(12). The asymmetry factor for the residual image.Col:(13). r−band apparent magnitude of the residual image. Col:(14). The flux ratio of residual and initial image within 3Re (Here Re

is the fitted effective radius from our model).

4.2 Dependence on environments

In the hierarchical structure formation model, clusters ofgalaxies form through mergers of sub-clusters and smallgroups; in the process central galaxies are expected to growtogether with clusters. This scenario suggests that the evo-lution of BCGs may depend on the environments. We thusexamine the relation between the fraction of BCGs involvedin major dry-mergers and the richness of the parent clusters.

Figure 5 shows the results. The fraction of mergingBCGs appears to increase as the clusters become richer. Asimple linear fit in the form of fraction = A log richness +Byields A = 0.11 ± 0.04, B = −0.12 ± 0.05 (shown as thesolid line), and so the significance is not very high due tothe large error bars. The higher frequency of more massivedry-mergers in richer clusters will lead to more luminous (or

c© 2009 RAS, MNRAS 000, 1–9

Major Dry-Mergers In BCGs 7

Figure 5. The top panel shows the major dry-merger fractionas a function of the richness of the parent clusters. Poisson er-rors are shown. The straight line shows the best regression fitfraction = (0.11 ± 0.04) log richness − (0.12 ± 0.05). The bottompanel shows the BCG luminosities versus the richness for 85 BCGs(open circles) in Paper I and 18 merging BCGs presented in this

work (solid circles).

more massive) BCGs. This is empirically seen in observa-tions, as illustrated in the bottom panel of Figure 5.

5 SUMMARY AND DISCUSSION

In this work, we have searched for BCGs involved in ma-jor dry-mergers in a sample of early-type BCGs selectedcarefully from C4 cluster catalogue (Miller et al. 2005). 18(∼3.5%) major dry-mergers involving BCGs have been iden-tified out of 515 clusters at 0.03 6 z 6 0.12. These majormergers are selected as pairs (or triples) with r-band magni-tude difference δmr < 1.5 and projected separation rp < 30

kpc, and showing signatures of interaction in the form ofsignificant asymmetry in residual images.

To double check whether our results are affected by thechoice of the cluster (C4) catalogue (based on DR2 of SDSS),we have also searched for major dry-mergers involving BCGsin the sample of 625 clusters refined by von der Linden etal. (2007) from the unpublished DR3 version of C4 clus-ters. Their sample selects preferentially early-type BCGsand discards clusters with very few member galaxies. Forthis sample, we find 27 (27/625∼4.3%) BCGs are involvedin major dry-mergers, entirely consistent with the fraction(3.5%) found here for the C4 cluster catalogue. It shouldbe emphasized that the goal of our method is to identifymerging pairs with strong interaction features and calculatetheir true faction in BCGs. They should be real, physicallybound pairs. However, some physical bound pairs may notyet show strong interaction signatures. Our estimated frac-tion should thus be considered as the lower limit of physicalpair fraction in BCGs.

From the identified major-merger candidates, we con-clude that a present-day BCG may have experienced0.6 (tmerge/0.3Gyr)−1 major dry-mergers and increasedtheir mass by 15% (tmerge/0.3Gyr)−1(fmass/0.5) on averagefrom z = 0.7. This fraction is comparable to the mass in-crease predicted by mostly minor mergers in semi-analyticalgalaxy formation models of De Lucia & Blaizot (2007), if allmass of companion galaxies are added to BCGs (fmass = 1)as De Lucia & Blaizot (2007) assumed. However, the massincrease we found here is due to major-mergers rather thanminor majors. Bernardi (2009) concluded that the evolu-tion of BCGs in the velocity dispersion-stellar mass relationfavors the growth of BCGs via many dry minor-mergers.However, the evidence is not very strong (see her Fig. 8). Amore detailed comparison between theory and observationsis desirable.

The major uncertainties in our estimates arise due tothe calibration of the merging timescale and the fraction ofthe satellite galaxy that will be added to the central host.For the former, we used the calibration of Kitzbichler &White (2008) derived from the Millennium simulation. How-ever, it remains to be seen whether their calibration applieswell to the dense environments such as clusters of galax-ies. The parameter fmass is also somewhat uncertain. Thestellar mass not accreted to the central host may form theenvelopes in cD galaxies and/or the intracluster light (Lin &Mohr 2004; Rines et al. 2007; Paper I and reference therein).We notice, however, that if all the mass is assembled intothe final BCGs, some of them may be too bright; the re-sulting luminosity function at the bright end may be incon-sistent with the observations which show little evolution atL > 2.5L∗ (e.g., Brown et al. 2007; Cool et al. 2008). Weconclude that a substantial fraction of the companion massmust be stripped, consistent with the presence of envelopesin giant cD galaxies (e.g., Paper I).

We have also checked the radio properties of BCGs toinvestigate whether major mergers trigger the phenomenonof active galactic nuclei. We cross-correlate our sample withthe radio data from the FIRST survey (Becker, White &Helfand 1995). There are 94 (∼18.3%) sources with radioemission out of 515 BCGs; these fractions are consistentwith the statistics of Best et al. (2007) and Croft, de Vries& Becker (2007). In comparison, there are 3 radio sources

c© 2009 RAS, MNRAS 000, 1–9

8 F. S. Liu et al.

(∼16.7%) out of 18 merging BCGs, i.e., there is no differenceon radio properties between major-merger and non-major-merger BCGs within the statistical uncertainties.

McIntosh et al. (2008) identified 38 major mergers ofred galaxies from 845 groups at z 6 0.12, out of which18 mergers (∼ 2.1%) are between a central galaxy and asatellite galaxy. Six (C1000, C1004, C1011, C1060, C1304,C3157) of our 18 mergers are also in their sample of 38 ma-jor mergers; all six are in the top ten most massive groupsamong their major mergers. In other words, we are probingthe most massive end of their SDSS “groups” in terms ofmass range from 1013.5M⊙ to 1015M⊙. They find that moremassive mergers tend to appear in richer systems, a conclu-sion confirmed by our studies. In addition, they find thatthe merger frequency increases as a function of halo mass,which may explain why their merger fraction (∼ 2.1%) isslightly lower than ours (3.5%). The mass increase rate wefind for the BCGs is roughly consistent with their results(2%-9% per Gyr under similar assumptions). We concludethat major dry-mergers in BCGs are not rare events in thelocal Universe and they are an important way to assembleluminous BCGs.

ACKNOWLEDGMENTS

We thank C. N. Hao, J. L. Han for useful discussions, andin particular A. von der Linden for sharing data and help-ful discussions. We acknowledge the anonymous referee for aconstructive report that improved the paper. This project issupported by the Doctoral Foundation of SYNU of China(054-55440105020), and by the NSF of China 10833006,10778622 and 973 Program No. 2007CB815405. SM ac-knowledges the Chinese National Science Foundation andTianjin Municipal Government for travel support. Fund-ing for the creation and distribution of the SDSS Archivehas been provided by the Alfred P. Sloan Foundation, theParticipating Institutions, the National Aeronautics andSpace Administration, the National Science Foundation,the U.S. Department of Energy, the Japanese Monbuka-gakusho, and the Max Planck Society. The SDSS Web siteis http://www.sdss.org/. The SDSS is managed by the As-trophysical Research Consortium (ARC) for the Participat-ing Institutions. The Participating Institutions are The Uni-versity of Chicago, Fermilab, the Institute for AdvancedStudy, the Japan Participation Group, The Johns Hop-kins University, the Korean Scientist Group, Los AlamosNational Laboratory, the Max-Planck-Institute for Astron-omy (MPIA), the Max-Planck-Institute for Astrophysics(MPA), New Mexico State University, University of Pitts-burgh, Princeton University, the United States Naval Ob-servatory, and the University of Washington.

REFERENCES

Becker R. H., White R. L., Helfand D. J. 1995, ApJ, 450,559Bell E. F. et al., 2006, ApJ, 640, 241Best P. N., von der Linden A., Kauffmann G., Heckman T.M., Kaiser, C. R.2007, MNRAS, 379, 894Bernardi M. et al., 2003, AJ, 125, 1849

Bernardi M., Hyde J. B., Sheth R. K., Miller C. J. NicholR. C., 2007, AJ, 133, 1741Bernardi M. 2009, MNRAS accepted, arXiv e-print(arXiv:0901:1318)Binney J., Tremaine S. 1987, in Galactic Dynamics (Prince-ton: Princeton University Press)Blanton M. R., Roweis S., 2007, AJ, 133, 734Brown M. J. I., Dey A., Jannuzi B. T., Brand K., BensonA. J., Brodwin M., Croton D. J., Eisenhardt P. R. 2007,ApJ, 654, 858Combes F., Rampazzo R., Bonfanti P. P., Pringniel P., Su-lentic J. W. 1995, A&A, 297, 37Conselice C. J., Bershady M. A., Jangren, A., 2000, ApJ,529, 886Conselice C. J., Bershady M. A., Dickinson M., PapovichC., 2003, AJ, 126, 1183Cool R. J. et al., 2008, 682, 919Croft S., de Vries W.; Becker R. H. 2007, ApJ, 667, 13De Lucia G., Blaizot J., 2007, MNRAS, 375, 2De Propris R., Liske J., Driver S. P., Allen P. D., CrossN. J. G., 2005, AJ, 130, 1516Gao L., Loeb A., Peebles P. J. E., White S. D. M., JenkinsA., 2004, ApJ, 614, 17Jeltema T. E. et al., 2007, ApJ, 658, 865Jones C., Forman W., 1984, ApJ, 276, 38Kartaltepe, J. S. et al., 2007, ApJS, 172, 320Khochfar S., Silk J., 2008, arXiv e-print (arXiv:0809.1734)Kitzbichler M. G., White S. D. M., 2008, MNRAS, 391,1489Lauer T. R., 1988, ApJ, 325, 49Lauer T. R. et al., 2007, ApJ, 662, 808Le Fevre O. et al., 2000, MNRAS, 311, 565Lin L. et al., 2004, ApJ, 617, L9Lin L. et al., 2008, ApJ, 681, 232Lin Y.-T., Mohr J. J., 2004, ApJ, 617, 879Lotz J. M. et al., 2008, ApJ, 672, 177Liu F. S., Xia, X. Y., Mao S., Wu H., Deng Z. G., 2008,MNRAS, 385, 23 (Paper I)Masjedi M. et al., 2006, ApJ, 644, 54McIntosh D. H., Guo Y., Hertzberg J., Katz N., Mo H. J.,van den Bosch F. C., Yang X., 2008, MNRAS, 388, 1537Merritt D., 1985, ApJ, 289, 18Miller C. J. et al., 2005, AJ, 130, 968Mulchaey J. S. et al., 2006, ApJ, 646, 133Naab T. et al., 2006, ApJ, 636, L81Patton D. R., Carlberg R. G., Marzke R. O., Pritchet C. J.,da Costa L. N., Pellegrini P. S., 2000, ApJ, 536, 153Ostriker J. P., Tremaine S., 1975, ApJ, 202, L113Patton D. R. et al., 2002, ApJ, 565, 208Peng C. Y., Ho L. C., Impey C. D., Rix H.-W., 2002, AJ,124, 266Postman M., Lauer T. R., 1995, ApJ, 440, 28Rines K., Finn R., Vikhlinin A., 2007, ApJ, 665, L9ix H. R., White S. D. M., 1989, MNRAS, 240, 941Ruszkowski M., Springel V., ApJ in press, arXiv e-print(arXiv:0902.0373)Sersic J. L., 1968, Atlas de Galaxias Australes (Cordoba,Argentina: Observatorio Astronomico)Smith G. P., Kneib J.-P., Smail I., Mazzotta P., EbelingH., Czoske O., 2005, MNRAS, 359, 417Stoughton C. et al., 2002, AJ, 123, 485Strateva I. et al., 2001, AJ, 122, 1861

c© 2009 RAS, MNRAS 000, 1–9

Major Dry-Mergers In BCGs 9

Tran K.-V. H, van Dokkum P. G., Franx M., IllingworthG. D., Kelson D. D., Schreiber N. M. F., 2005, ApJ, 627,L25Tran K.-V. H., Moustakas J., Gonzalez A. H., Bai L., Zarit-sky D., Kautsch S. J., 2008, ApJ, 683, L17Tremaine S., 1990, in Dynamics and Interactions of Galax-ies ed. R. Wielen ( Berlin: Springer), 394Vale A., Ostriker J. P., 2008, MNRAS, 383, 355van Dokkum P. G., Franx M., Fabricant D., Kelson D. D.,Illingworth G. D. 1999, ApJ, 520, L95von der Linden A., Best P. N., Kauffmann G., WhiteS. D. M., 2007, MNRAS, 379, 867Wen Z. L., Liu F. S., Han J. L., 2009, ApJ, 692, 511Whiley I. M. et al., 2008, MNRAS, 387, 1253White S. D. M. 1976, MNRAS, 174, 19Yang X., Mo H. J., van den Bosch F. C., Jing Y. P., 2005,MNRAS, 356, 1293Yang X., Mo H. J., van den Bosch F. C., 2009, ApJ, 693,830

c© 2009 RAS, MNRAS 000, 1–9